Do accountability mechanisms in safety nets improve access to social services? The case of Brazil s Bolsa Família

Transcription

1 Do accountability mechanisms in safety nets improve access to social services? The case of Brazil s Bolsa Família International Institute for Democracy and Electoral Assistance (International IDEA) Anja Linder* 1 * This paper was prepared by Anja Linder, and was a part of International IDEA s Democracy and Development programme work in This document was selected as a contribution to stimulate debate on and increase knowledge about the impact of democratic accountability on services.

2 International Institute for Democracy and Electoral Assistance 2012 The International Institute for Democracy and Electoral Assistance (International IDEA) publications are independent of specific national or political interest. Views expressed in this publication do not necessarily represent the views of International IDEA, its Board or its Council members. 2

3 Contents 1. INTRODUCTION SAFETY NETS, DEMOCRACY AND DECENTRALIZATION ACCOUNTABILITY ARRANGEMENTS IN SAFETY NETS THE IMPACTS OF ACCOUNTABILITY MECHANISMS ON DEVELOPMENT OUTCOMES CONCLUSIONS REFERENCES

4 1. INTRODUCTION Are accountability mechanisms effective at improving service delivery for the poor? Improving public service delivery and strengthening the voice of the poor in affecting the pattern of provision and quality of services are central aspects of the development of modern democracies and equitable societies. The availability of mechanisms for ensuring that policymakers and other actors are held accountable for the delivery of quality public services is a crucial aspect of the process. An accountability mechanism is here defined as an institution or policy which enables citizens to voice preferences, complaints and concerns with the ultimate aim of holding public actors to account. Conditional cash transfers (CCTs) have been introduced in many countries across the world as a means of tackling household poverty, both directly through the transfer of money to the poor and indirectly by making transfers conditional on strengthening human capital within families, thereby preventing the inter-generational transfer of poverty. CCTs also provide opportunities to directly address service delivery to the poor and focus on the demand side of the relationship. There is evidence that the introduction of democratic accountability mechanisms into social service delivery has led to improvements in both the take-up of public services and development outcomes. According to Ferreira and Robalino (2010), CCTs in Latin America have helped increase school enrolment, bringing poor children to school, and have increased demand for health services among the poor. Overall, they conclude that: CCTs have transformed the social assistance landscape in Latin America in large part because they greatly enhanced the State s capacity to target resources effectively to the poorest people in society. The bottom fifth of the income distribution had been effectively excluded from any serious state assistance throughout Latin America s history, and this only started to change with the introduction of [non-contributory social insurance] and CCTs in the 1990s. (Ferreira and Robalino 2010: 25). The introduction of CCTs in developing countries has the potential to change accountability relations in social service delivery by introducing rights-based approaches to safety nets. 2 Utilizing the analytical framework of the 2004 World Development Report, which emphasizes the need for accountability in public service delivery, this paper analyses the accountability mechanisms introduced in the Bolsa Família Program (BFP) in Brazil in order to assess their effectiveness at improving access to social services for the poor. The analysis reveals important potential for strengthening accountability through two main mechanisms: the BFP s conditionalities and Social Control Councils (SCCs). Despite the lacklustre achievements of the social control mechanisms, the BFP has managed to improve take-up of social services and has improved social development outcomes. This paper provides a brief description of the BFP and the decentralized governance context in which it functions. It describes the specific accountability mechanisms of the BFP. A number of indicators are proposed to measure the performance of programme accountability mechanisms, and some general conclusions are drawn on the benefits and challenges of accountability mechanisms when seeking to improve the delivery of social services. 4

5 2. SAFETY NETS, DEMOCRACY AND DECENTRALIZATION 2.1 BRAZIL AND THE BOLSA FAMÍLIA PROGRAMME Brazil has a long history of high levels of poverty and inequality. Its level of inequality ranks among the highest in the world, reflected by a Gini coefficient of about 0.6 since the 1970s. The BFP was created in 2003 in a context of long-standing and substantial popular support for public spending on measures to combat poverty and inequality, even among the country s elite. In a survey of 320 randomly selected individuals from four sectors of Brazil s elite (political, governmental, private sector, and union leaders), close to half ranked issues such as poverty, inequality and access to education as the main obstacles to greater democracy in Brazil, and close to 40 per cent ranked them as the priority issues facing the nation. 3 The long debate about social justice and the need to provide a minimum income to the poorest, which preceded the creation of the first cash transfer programmes and which continues today, shows that well-directed poverty programmes continue to enjoy broad societal support. Concerns about social injustice are reflected in Brazil s 1988 Constitution and have underpinned its transition to democracy. The constitution places great emphasis on poverty reduction and the creation of a more just and equitable society. In the fundamental principles of the federation, the constitution includes the elimination of poverty and marginalization, and a reduction in social and regional inequality. According to Cunha and Câmara (2008:1), the constitution represents the crowning of an historical process of revising political institutions, of redemocratization, of reassessing relations between the state and society, and of guaranteeing social rights. Moreover, it establishes responsibilities on the part of the state to strengthen democratic liberties and establish new means of social participation, including the creation of specific channels for social participation in the control of public management. The constitution also establishes a legal foundation for social assistance as guaranteed rights for the needy, and thus an obligation on the state to provide health and education services, among others, the free access to which is established as a basic right of all citizens. In order to guarantee continued focus on these issues, constitutional minimum spending levels as set percentages of the federal budget were established. In terms of subnational spending, the 1988 Constitution instituted an earmark of 25 per cent of municipal spending for education. Subsequent amendments to the constitution added earmarks of 15 per cent of municipal spending to health, and 25 per cent and 12 per cent, respectively, of state spending for education and health (Souza 2009). Although the constitution establishes the right of access to basic services for all, it does not necessarily guarantee the take-up of these rights by the poor. The direct and indirect, or opportunity, costs of these services mean that many are still excluded. Cash transfers conditioned on the use of public social services, first introduced in Brazil at the municipal level in 1995, thus serve as an instrument not only to provide a minimum income for the poor, but also to encourage and require take-up of basic social services. The BFP was created in October 2003, through the merger of four pre-existing cash transfer programmes in an effort to improve the efficiency and coherence of the social safety net and to scale-up assistance to provide universal coverage for Brazil s poor. The creation of the BFP should be seen in the context of long-standing pent-up social demands and a political consensus that these demands needed to be met for the good of the country as a whole. 5

6 Recurring financial crises, which hit the poor especially hard, and the gradual strengthening of the political left, with its calls for social justice, culminated in the election of President Luiz Inácio Lula da Silva, who provided a strong voice for the cause of improving the social safety net and introducing the right to a minimum income for the poor. Indeed, the reelection of Lula in 2006 was based largely on the successes of the BFP. This, in some sense, represented a paradigm shift, with a government being judged on the efficiency of a social policy rather than rewarded for clientelist practices distributing social programme benefits in return for votes, as had often been the case in the past. The BFP is the largest CCT programme in the developing world, covering over 11 million families across the whole country. Like other CCTs, the BFP seeks to help reduce current poverty and inequality, by providing a minimum level of income for extremely poor families, and to break the intergenerational transmission of poverty by making these transfers conditional on beneficiary compliance with human capital requirements in the areas of health and education. One of the main aims of the programme, as specified by the laws which created it, is to promote access to the network of public services, especially in health, education and social assistance. 4 The BFP also seeks to help empower BFP beneficiaries by linking them to complementary services. The BFP provides a fixed transfer of BRL 70 (USD 42) per month to all families in extreme poverty, officially defined in 2011 as families with a per capita income of up to BRL 70 per month, regardless of their demographic composition. Moderately poor families, that is, families with a monthly per capita income of up to BRL 140 (USD 85), receive monthly transfers of BRL 32 (USD 19) per child up to age 15, for a maximum of five children. Total family cash transfers range from a monthly minimum of BRL 32 to a maximum of BRL 306 (USD ) depending on family composition and the level of income. 5 Families with children aged 16 to 17 who are still in school are given BRL 38 (USD 23) per child for a maximum of two children. 6 Receipt of transfers is made conditional on accessing education and health services. In education, children and young people aged 7 to 17 must enrol in school and attend regularly (a minimum attendance level of 85%). In terms of health, children are required to have their vaccinations up to date and must undergo regular health check-ups and growth monitoring. Pregnant women must attend pre- and post-natal check-ups and participate in educational health and nutrition seminars. As with other CCTs, the conditions for receiving BFP transfers serve to promote access to basic social services for the poor. Public health and education services are widely available in Brazil, but many poor families do not access these services mainly due to the high direct and opportunity costs. The BFP helps to increase access to these services by providing families with cash to offset the costs of access, and conditions the receipt of transfers on the effective use of health and education services. The focus is very much on helping families to overcome barriers to the use of public social services. Lack of compliance with the conditionalities is viewed as a sign of vulnerability and of the need for further assistance in order to improve living conditions and ensure sustainable social inclusion (Lindert et al. 2007). Only after repeated rounds of non-compliance are families penalized and potentially excluded from the transfer programme. 6

7 2.2. DECENTRALIZATION AND THE DELIVERY OF SOCIAL SERVICES IN BRAZIL The BFP inserts itself into Brazil s highly decentralized context of policy implementation. The 1988 Constitution established shared responsibility for most social policies between the three levels of government: the federal, the state (26 states plus the federal district of Brasilia) and the municipal (5564 municipalities). According to Arrteche (2004), any federal entity has the constitutional right to implement programmes in areas such as health, education and social assistance, among others. Furthermore, the division of responsibilities between levels of government varies within each specific area of social policy, with some being more centralized than others. From the mid-1990s, local governments became the main providers of basic health and education services, but such service provision is based on federal resources and spending mandates in order to ensure minimum levels of service provision (Souza 2005). Decentralization poses particular challenges for the implementation of federal social programmes. Although managed at the federal level, by the Ministry of Social Development (MDS), many aspects of BFP implementation are carried out by Brazil s 5564 municipalities. The constitution establishes joint responsibility between the three levels of government in the areas of social assistance and fighting poverty and inequality, but the BFP was created by regular law and does not have constitutional status. The challenges of decentralization include: the principal-agent dilemma involved in executing federal programmes through autonomous municipalities; the heterogeneity in municipal administrative and financial capacities for programme implementation; and the interactions between the BFP and local social programmes. Formal mechanisms for programme implementation have been developed, alongside performance-based management tools and incentives, in order to overcome some of these challenges. The attempts to overcome these challenges have spawned a plethora of operational solutions. For example, formal joint management agreements have been signed by the MDS and all municipalities in order to formalize municipal roles and responsibilities in programme implementation, and to establish minimum standards for programme operation. The quality of municipal implementation is assessed using an Index of Decentralized Management based on which financial incentives are awarded and training and capacity building is provided where quality is low (Cunha and Câmara 2008). To avoid duplication with local social programmes, joint cooperation agreements provide for the vertical integration of such programmes with the BFP. In addition, the annual BFP Innovations Award promotes the sharing of experiences and learning (Lindert et al. 2007). Agreements between the federal government and the states and municipalities have thus formalized responsibilities at each level. Municipalities are responsible for the ongoing implementation of the programme. They have to maintain a local coordinator for the programme and register potential beneficiaries in the national database for social transfer programmes (Cadastro Único). They also monitor health and education conditionalities and consolidate the associated information. Municipalities help to channel BFP beneficiaries towards other complementary services and also have a duty to establish Social Control Councils. State governments provide technical support and training to municipalities. They also provide basic and complementary services, and are responsible for providing identification documentation for all families in the Cadastro Único registered in their jurisdiction. The ministries of health and education are responsible for establishing technical and operational 7

8 guidelines regarding school attendance and health conditionalities, and are also responsible for promoting training in monitoring conditionalities. In addition, they are in charge of consolidating conditionality compliance information and reporting this information to the MDS. Three controls agencies the General Comptroller s Office (Controladoria Geral da União, CGU), the Federal Audits Court (Tribunal de Contas da União, TCU) and the Office of the Public Prosecutor (Ministério Público, MP) are responsible for formal oversight and control of the BFP. 3. ACCOUNTABILITY ARRANGEMENTS IN SAFETY NETS The analysis of BFP accountability mechanisms is based partly on the framework presented in the 2004 World Development Report: Making Services Work for Poor People (World Bank 2003), which stresses the need for accountability in public service delivery. Within this context, the role of the BFP itself as well as various aspects of programme implementation are analysed in order to establish how the BFP ensures accountability to citizens/beneficiaries and the limits to such accountability ANALYTICAL FRAMEWORK: CITIZENS AS CLIENTS The 2004 World Development Report underscores the need for accountability in service delivery in order to strengthen the incentives of service providers to improve access to and the quality of public services, particularly for poor people. The report emphasizes the need to strengthen accountability in three key relationships in the service delivery chain. First, citizens can resort to the long route to accountability, using voice and politics in the citizen politician/policymaker relationship. By voicing concerns to politicians through the democratic process, citizens can provide incentives for their elected officials to hold service providers accountable for the quality and level of service provision. Where the long route to accountability is dysfunctional, strengthening client power in the citizen/client service provider relationship can give citizens, as clients, greater power over providers by shifting decisions and power directly to citizens or communities. This strengthens the short route to accountability and can help clients directly affect and improve the supply and quality of local services. The policymaker provider relationship is characterized as a compact, based on which policymakers provide resources and delegate powers to service providers, monitor the performance of the providers and subsequently provide rewards or impose sanctions on providers based on their performance in the provision of public services. The report highlights the weakness of government incentives to improve service delivery for the poor. Policymakers often have few means of holding service providers accountable for the services they provide. Moreover, this problem is particularly acute when the state is the service provider. Greater emphasis has therefore been placed on strengthening client power and the short route to accountability. Policies such as decentralization and citizen participation have been used to improve voice in order to improve provider responsiveness (Joshi 2008). As is discussed above, Brazil s governance framework is highly decentralized. Various mechanisms of social participation in the policy process, such as participatory budgeting, have been implemented in the recent past. The social participation feature of the BFP is discussed below. 8

9 CCTs such as the BFP could help strengthen the short route to accountability by increasing flows of information to and the power of beneficiaries. BFP beneficiaries poor families, many of which have previously been excluded from basic public services are now required to access services such as education and health in order to receive the BFP transfer. As they increase their use of services, they acquire a new identity as clients of public services, with a voice and potentially a certain measure of power over providers. This shift in identity has relevance both externally, vis-à-vis others, such as service providers and policymakers, and internally, vis-à-vis the beneficiaries themselves, enhancing their self-esteem and strengthening their ability to exercise voice. With the right mechanisms of information, social participation and oversight, the BFP could help shift governance arrangements through the increased involvement of citizens in various aspects of programme implementation and monitoring as well as the delivery of public services linked to the programme, as is described below. 3.2 MECHANISMS FOR ACCOUNTABILITY The BFP has, through the monitoring of conditionalities, provided a new means of enhancing access to public services for the poor. The programme has therefore, in some sense, functioned as an accountability mechanism in itself. As such, it has managed to strengthen service take-up while also positively affecting development outcomes. The central instrument for accountability created within the BFP is the provision for social control and oversight carried out through the Social Control Councils. By including beneficiary representatives on the councils, this mechanism should enable poor people to participate in the monitoring of service provision at the local level and to strengthen incentives for service providers to serve the poor, thus enhancing client power and strengthening the short route to accountability. 1. The Bolsa Família Programme as an accountability mechanism As is noted above, the 1988 Constitution introduced guarantees of universal rights of access to social services such as health and education for all and placed the responsibility for the provision of such services on the public authorities. Constitutional mandates notwithstanding, access to and take-up of services by the poorest and most vulnerable were not guaranteed and many are excluded due to the direct and opportunity costs associated with use of the services. For instance, in terms of public and free education services, direct costs such as books and other supplies as well as opportunity costs, such as the time spent in school rather than working to earn money for the family, serve as barriers to the take-up of this service for many poor families (Glewwe and Kassouf 2010). Similarly, using free public health services entails direct costs such as travelling to the health clinics and indirect costs linked to the opportunity costs of alternative uses of travel time. The BFP and its programme conditionalities, which require families to access health and education services, could be viewed as strengthening citizens rights and reinforcing the right of access to services for the poorest. More specifically, the existence of conditionalities in terms of accessing education and health services elicit demand-side responses on the part of the beneficiaries, and also potential supply-side responses on the part of public authorities. On the demand side, CCTs can help bring services to the poor through an income effect, raising disposable income and thus off-setting the costs of accessing services, as well as a price-effect, as conditionalities make it costlier not to access the services associated with the programme. 9

10 Because compliance with conditionalities can only be demanded to the extent that the state provides an adequate supply of the services and guarantees accessibility for the poorest, CCTs could also generate a supply-side response. The conditionalities then become a means of ensuring that public authorities actively seek out the poorest where they do not present themselves to take up services. The right to a basic income thus reinforces the right to education and health (Cunha and Câmara 2008, Rocha 2008). According to Rocha (2008), the BFP is not a right per se, but it is a public policy which aims to ensure access to fundamental rights. The monitoring of families thus represents an important social policy and a democratic channel allowing the state accurately to detect individual vulnerabilities in order to direct social services and assistance where needed. The conditionalities thus provide a mechanism for connecting the poor to the services constitutional rights to which they are entitled but which they have often been unable to benefit from due partly to cost barriers. 2. Mechanisms for social control of the BFP Potentially the most important accountability mechanism of the BFP is its provision for social control and oversight over various aspects of programme implementation through the Social Control Councils. The idea that material improvements for ordinary citizens represent favours to be delivered by elite political mediators was eventually challenged by social movements, independent associations and ordinary citizens (Avritzer 2002:5). According to Hevia (2008: 3), the creation and strengthening of social participation in Brazilian social policy following the process of democratization and the adoption of the 1988 constitution represents a crucial institutional innovation for [Brazilian] democracy. The importance of participatory bodies, such as SCCs, which exist in various social policy areas, has made the Brazilian model one of the world paradigms of participation and social control (Hevia 2008, Avritzer 2002). Social control exists in various areas of Brazilian social policy. As part of the BFP s administrative responsibilities, municipalities are obliged to establish local SCCs. SCCs can be specifically created for the control and oversight of the BFP, or existing councils in other social policy areas can be used for the BFP. According to the MDS (n.d.), just over 60 per cent of SCCs were specifically created for the BFP. Of the remaining, the vast majority (over 90%) of the existing councils used for BFP were Social Assistance Councils. SCCs should have representatives of local authorities and civil society in equal numbers, following the participatory management model that governs Brazil s social protection policy (Hevia 2010, 2009). SCCs must also include programme beneficiaries, and practitioners in the areas of social assistance, health, education and food security and, where appropriate, should include children and adolescents. The representatives from civil society must be selected independently of the local government and governing authorities. 7 According to MDS (n.d.), 52 per cent of the nearly members of Social Control Councils are representatives of civil society while 48 per cent represent local government. Furthermore, the majority of the members are women (64% versus 36% men according to MDS data), about 35 per cent have a high school diploma and just over 30 per cent have completed undergraduate studies. Around 13 per cent of members had not completed their elementary schooling. The SCCs have an important role to play in the enforcement, monitoring and evaluation of the BFP. More specifically, SCCs check that the poorest areas of the municipality are being 10

11 prioritized in cadastro registration (see below). They verify the local inclusion of poor and extremely poor families in the cadastro registry and whether those families are then actually included in the BFP. They are also in charge of periodically evaluating the local list of beneficiaries to ensure it reflects local realities and includes the truly poor, thus minimizing errors both of inclusion and exclusion. SCCs also monitor the system of benefits management to verify the suspension, blocking, cancellation or reactivation of programme benefits. SCCs are also responsible for monitoring the local provision of services to ensure that municipalities provide the education and health services needed for beneficiaries to be able to comply with the conditions of receiving the transfer payments. In addition, they should check whether the municipality has the mechanisms needed to monitor families that for some reason have not been able to comply with these conditions in order to identify possible solutions (MDS n.d.). Moreover, they are supposed to promote community participation in the oversight of BFP implementation and contribute to the development of public information campaigns concerning the BFP (Lindert et al. 2007, Hevia 2008). 8 Municipalities are responsible for setting up the SCCs and providing them with sufficient and timely information to enable their monitoring of local programme implementation. They are also required to disseminate information to the local population concerning the role and relevance of the SCC (MDS n.d.). In cases in which SCCs detect any irregularities in local programme management, they should first warn the municipal programme management unit in order to find an adequate solution. Where no solution can be found, SCCs should inform the MDS. In certain more serious cases of irregularity or abuse, the MDS should inform the public oversight network, comprised of the three federal control agencies with varying levels of power of control and sanction listed above (see also below). 9 This is a hypothetical procedure as in reality, each SCC is autonomous and there is significant variation in terms of procedures and specific actions taken in case of irregularities. The SCCs represent a potentially very important instrument for social representation and accountability, but there is considerable doubt concerning their functionality and impact. Lindert et al. (2007: 80) cite municipal survey data from 2004 which show that SCCs can have a positive impact on the quality of programme implementation when they work. However, the same survey discovered weaknesses in SCC functioning and composition. For instance, many SCCs did not meet regularly, and over 50 per cent of the SCCs surveyed did not have an up-to-date list of beneficiaries. There were also reports that 60 per cent of the SCCs did not regularly visit schools or health facilities to oversee the monitoring of conditionalities. Moreover, although SCCs are supposed to be composed of equal numbers of representatives of civil society and local government, the survey showed that over 70 per cent of members were supporters of the local mayor, and 7 per cent were family members of the mayor. Operational audit data from the CGU from highlight similar weaknesses, with SCC membership skewed towards local authorities, and municipalities lacking adequate information to perform their social oversight function (Lindert et al. 2007). Furthermore, although BFP beneficiaries should be represented on all SCCs, this is often not the case. In fact, many beneficiaries are unaware of the existence of SCCs. It is difficult to imagine true social oversight and control under such circumstances. Where the SCC is not truly independent of local government, any powers it might potentially have could easily be neutralized in case of a possible conflict with municipal government interests. Many observers agree with this criticism, claiming that social control of the BFP is largely ineffective. SCCs lack autonomy and beneficiaries have only weak powers to represent themselves and their interests. While the SCCs have sufficient autonomy on paper, and are 11

12 able to establish their own rules of operation, beneficiaries are under-represented and the SCCs themselves lack the minimum powers, resources and control necessary to enable the empowerment of beneficiaries and the exercise of strong social control (Hevia 2009). Furthermore, the lack of knowledge by beneficiaries of the SCCs has serious implications, limiting the possibility that SCCs will truly function for and represent beneficiaries (Hevia 2008). Contacts with the MDS have confirmed this assessment. The MDS concedes that SCCs do not have much power to act on or resolve specific problems or irregularities. It states that the SCCs are more oriented to simply monitoring the BFP and to being a point of contact for beneficiaries. However, the fact that many beneficiaries and the public in general are unaware of the existence as well as the purpose of SCCs casts serious doubt on this function as well. 10 This lack of power resources and of opportunities for direct intervention set the BFP SCCs apart from the so-called coordinating councils present in other social policy areas. Councillors in other areas, such as the management of the Single System for Social Assistance (SUAS), have veto power over municipal resources and can intervene with the highest policymaking authority in the area the national conference on social assistance. The BFP Social Control Councils have none of these powers and are limited to monitoring and oversight of the programme through meetings with the municipal programme manager (Hevia 2009). These shortcomings seriously limit the potential of BFP SCCs to function as intended. The SCCs are a potentially crucial mechanism for accountability within the BFP and could be a potent tool for beneficiary representation and defence in case of abuses on the part of the municipal authorities. However, in their current form, they are unlikely to exercise their functions in an effective manner or to undertake vital social control and oversight of the programme. 3. Other programme mechanisms to promote accountability 11 In addition to the SCCs, the BFP also has other mechanisms for enhancing transparency, credibility and, ultimately, accountability. Such means include a public network for oversight and control as well as telephone hotlines and an ombudsman to provide information and channel complaints. The three federal control agencies the CGU, the TCU and the MP and their subnational counterparts are responsible for formal oversight and control of the BFP. Together, these three agencies jointly form what is known as the official oversight and control network (rede de fiscalização) for the BFP. These entities work in close collaboration with the MDS to ensure overall control of the programme. The General Comptroller s Office (Controladoria Geral da União) is the internal control organ of the federal executive. It performs random-sample operational audits, annual financial audits and case investigations. In the context of the operational audits, the CGU verifies eligibility and cadastro data for a random sample of beneficiaries in each municipality, as well as monitoring compliance with conditionalities and monitoring implementation processes, including SCCs. The reports from such audits are shared with the MDS as well as the other two agencies in the federal control network, the federal Congress and the Police. The CGU also conducts case investigations in the municipalities for each operational audit, when the MDS forwards complaints received or irregularities are detected for a particular sample of municipalities. The CGU publishes BFP beneficiary lists on the government transparency portal website as public information. 12

13 The Federal Audit Court (Tribunal de Contas da União) provides the MDS with information, methodologies and lessons learned in order to improve monitoring of the BFP. The TCU has carried out implementation evaluations of the Cadastro Único as well as of the BFP itself. All activities undertaken by the TCU in this context aim to develop instruments to help improve the monitoring of programme funds. The Public Prosecutor s Office (Ministério Público) seeks to guarantee respect for the constitutional rights of individuals. It investigates cases which are usually generated by individual complaints or press reports. In the BFP context, the MP investigates allegations of fraud in beneficiary registration and assesses compliance with programme conditionalities. The MP can request registry lists from municipalities to verify the quality and accuracy of the targeting of the registry system or to detect potential irregularities, such as suspicions of political manipulation of the programme in election years. The MP can also propose sanctions in case of irregularities. In addition to the public control network, telephone hotlines are available to programme beneficiaries, programme managers and the general public. These hotlines serve to resolve questions or complaints in order to enhance transparency and ensure public access to programme information. The MDS Central Hotline responds to inquiries from beneficiaries and programme managers as well as to general complaints from the public through an Ombudsman (Ouvidoria). The hotline is linked directly to the office of the Executive Secretariat of the MDS. The hotline is staffed by generalists, who respond to general questions from beneficiaries; by specialists, who respond to questions from local BFP managers; and by representatives of the ombudsman. Most questions concern payment, blocked benefits and the process of eligibility selection. The ombudsman responds to suggestions, complaints and formal accusations from the general public. Calls to the ombudsman generally concern the quality of customer service in municipalities and at payment centres, questions about where to find information on the BFP, requests for information about blocked or cancelled BFP benefits and requests for information about how to receive the BFP benefit. The Bolsa Família secretariat within the MDS also operates a specialized hotline for inquiries from local BFP managers, in order to respond to questions concerning interpretation of the legal and normative framework of the BFP, management information systems and general questions about how best to provide services to BFP families. There is also a hotline for the programme payments agent, the Caixa Central Hotline. It has seven call centres, which respond to queries from beneficiaries and programme managers, mostly concerning payments and payment cards. All these mechanisms serve to enhance transparency and to minimize opportunities for fraud and abuse. This is important in terms of both providing general and specific information to beneficiaries and citizens at large, and anchoring political support for the programme. 4. THE IMPACTS OF ACCOUNTABILITY MECHANISMS ON DEVELOPMENT OUTCOMES As is discussed above, CCTs can help to strengthen the accountability of both the state and service providers to beneficiaries and citizens more generally. This can help enhance service 13

14 provision and benefit the poor. A central aspect of the BFP is the fact that many programme features were specifically designed to depoliticize it and avoid social and political intermediation in programme implementation. This was done to limit clientelism and opportunities for corrupt practices, and in these respects has been largely successful. Thus far, the BFP has obtained good results in terms of reaching the poor, through good targeting, and enhancing development outcomes. Nevertheless, certain constraints still limit its impacts. Stipulations for social control and oversight through the SCCs provide a potentially very important mechanism for ensuring social accountability. Strengthening mechanisms for social participation and oversight could be a potent tool for putting public pressure on policymakers to serve the poor RELATIONS BETWEEN ACTORS IN PROGRAMME IMPLEMENTATION Traditionally, social programmes have often been used for clientelist purposes, providing social benefits in return for political support. Past social policies in Latin America provide ample proof of this. According to Avritzer (2002: 7), clientelism became one of the principal ways of building political majorities [in Latin America]. Beneficiaries of particular social programmes often rely on local authorities for programme registration and the payment of benefits, providing room for political manipulation. This has increased the power of authorities over beneficiaries, who by definition are poor and often illiterate and generally less able to claim their rights and defend themselves against abuses. The BFP was specifically designed to circumvent political actors, particularly at the local level, in order to avoid clientelist practices. The whole process of programme management and implementation has thus been developed in order to minimize direct contact between beneficiaries and programme officials, and to make the processes of targeting, beneficiary selection and transfer payments as transparent as possible and based on clear, objective criteria rather than on the discretionary decisions of politicians or programme managers. A discussion of the various steps involved in BFP implementation, as well as of the relations among the actors involved in each step, is set out below. Programme targeting Programme targeting is carried out in a two-level process. The federal government first allocates programme beneficiary quotas to individual municipalities by comparing poverty estimates, based on national household surveys and census figures, with eligibility criteria for entry into the BFP. These criteria are established at the central level by the MDS (Lindert et al. 2007). This step mainly concerns establishment of municipal programme quotas, that is, the maximum number of poor families that can be included in the BFP in each municipality, at the federal level. Information and registration In conjunction with this, municipalities are responsible for keeping a list of poor families in order to identify beneficiaries for all the social assistance programmes. These lists make up the Cadastro Único or unique register for social programmes. Municipalities use spatial maps of poverty and vulnerability as well as other indicators of living standards to identify and target geographic concentrations of poor families in their jurisdiction. The municipalities are responsible for distributing information about the BFP and about the need to register in the Cadastro Único in order to be considered for inclusion in the BFP. They also collect data on potential beneficiaries, based on a standard federal family questionnaire, and draw up the 14

15 local Cadastro Único registry, from which beneficiaries are eventually selected (this process is described below). Lindert et al. (2007) note that there is considerable room for local adaptation in this step, resulting in significant variation across municipalities concerning the methods for implementing Cadastro Único registration. Municipalities can register all poor families with a per-capita family income below a certain level, which is higher than the cut-off level for eligibility for the BFP, in the Cadastro Único. 12 Families must be informed that registry in the Cadastro Único does not guarantee receipt of the BFP transfer. There is thus a separation between the local registration of potential beneficiaries and the federal establishment of local quotas for receipt of programme transfers, as maximum quotas are established for the number of local BFP beneficiaries but not for local registry in the Cadastro Único. This separation has several advantages, according to Lindert et al. (2007). Such advantages include: increased transparency regarding which households are interviewed for potential programme inclusion as the Cadastro Único register should include all poor families in the municipality; reduced scope for manipulation by local politicians seeking votes by registering households with a certain party affiliation to the detriment of others; a fairer and more transparent process of registration. Because all poor families can be registered, it is less likely that the poorest and most vulnerable will be excluded, who are generally less well informed or less well connected, and would tend to be penalized in a more competitive registry environment. However, separating the centrally determined programme quotas from the locally established poverty registry does not entirely eliminate the potential for politically motivated registration in the Cadastro Único at the local level. This is particularly the case because, in the words of Hevia (2008), local mayors hold the keys to entry to the BFP through their role in applying the Cadastro Único. Lindert et al. (2007) nevertheless argue that the risks of political manipulation are small due to the technical design of the programme, coupled with extensive provisions for monitoring and oversight. Indeed, BFP beneficiary databases are public and the SCCs, in which beneficiaries should also participate, help to oversee programme implementation. These factors all combine to provide incentives for municipalities to focus their limited beneficiary slots on those who are truly poor. Social oversight does in fact appear to be effective at exposing cases of inclusion of the non-poor into the programme, which is politically highly unpopular. Beneficiary selection The selection of beneficiaries to receive BFP transfers is finally carried out at the federal level by the MDS. The application of programme eligibility criteria to family data is carried out automatically using the Cadastro Único software, which compares self-reported incomes to the official eligibility thresholds. Through this process, families are prioritized and benefits are assigned according to income and family composition. The MDS then runs additional consistency checks to verify information and finalize the beneficiary list. The selection of beneficiaries was centralized under the BFP in contrast to the practices used for existing cash transfer programmes. This centralization has helped enhance coherence and consistency, but was also done with the explicit objective of reducing the scope for clientelist uses of federal funds by local authorities (Lindert et al. 2007). 15

16 The crucial step of selecting which families receive the benefit is thus largely automated and does not involve any direct contact between potential beneficiaries and programme officials. Instead, selection is made on the basis of objective eligibility criteria and data on family income. Continual verification of data as well as social oversight, based partly on public beneficiary lists, help to ensure the transfers are reaching the true target audience. Transfer payments Benefit transfers to beneficiaries are carried out entirely without intermediaries at the local level. Instead, payments are made through the banking system. Beneficiaries 13 receive an electronic benefits card through the postal system, and then choose and register a personal pin code through which they can retrieve their money at a variety of payment points throughout the country. Using the banking system for transfer payments has several advantages. It helps increase transparency and promotes efficiency by taking advantage of an extensive existing banking infrastructure. It also reduces the scope for clientelism by cutting out state and local public intermediaries from the payment process. The absence of intermediaries in the payment process represents a radical change compared with past practices of political clientelism, where programme handouts were strongly associated with specific politicians or political parties (Cunha and Câmara 2008). Monitoring conditionalities and benefits management Monitoring compliance with conditionalities implies first of all providing information to the public about the conditionalities and about sanctions for non-compliance. BFP conditionalities are widely publicized using television, radio, billboard and newspaper campaigns by both the federal and the subnational governments. All beneficiaries also receive information booklets about their responsibility to comply with the health and education conditionalities. Daily data on school attendance are collected by the schools and transmitted to the municipal education secretariat on a regular basis. Data are sent to the Education Ministry, which has formal responsibility for overseeing the monitoring of education conditionalities, on a monthly basis. Data are also shared bi-monthly with the MDS, which analyses the data and takes disciplinary action where warranted. 14 Compliance with health conditionalities depends greatly on access to services. In municipalities where the Family Health Programme (PSF) exists, beneficiaries should receive monthly visits from health workers who offer the services needed for compliance with the conditionalities. In such cases, the conditionalities serve to promote a supply-side response by getting services to families rather than punishing for non-compliance. In municipalities where the PSF does not operate, families have to seek care in hospitals or smaller health centres, which requires a greater level of effort on the part of beneficiaries. In both cases, data on compliance with vaccination and other requirements are gathered by the health workers. Consolidation of such data is less frequent than is the case for education, and information can be gathered at multiple points of service for any single beneficiary. Information is regularly transmitted to the municipality, which transmits it to the Ministry of Health twice a year. The Ministry of Health, which has formal responsibility for oversight of the health conditionalities, consolidates national data and transmits them to the MDS. In addition to sanctions for non-compliance with programme conditionalities, beneficiaries may have their transfers blocked, suspended or cancelled due to changes or irregularities in their Cadastro registry information, duplication of BFP payments with payments from the Eradication of Child Labour Programme (PETI) or by court order. Only the MDS can block 16

17 benefits due to non-compliance with conditionalities, but municipalities can block benefits for the above reasons. In such cases, the legally responsible BFP beneficiary can request a review of sanctions by submitting an appeal to the municipal BFP coordinators. The municipality is then responsible for investigating the case and responding to the beneficiary within 15 days. The design of the BFP has thus successfully limited social intermediation through social organizations and political intermediation through the participation of political parties or actors in each step of the implementation process. While social and political intermediation in programme administration and management has been eliminated, no new channels for intermediation were created for the purpose of the programme. This sets the BFP apart from other CCT programmes, such as Mexico s Oportunidades. Instead, institutional channels of communication are used, such as schools, the mass media, the postal system, poverty indicators and the banking system (Hevia 2008). All this has eliminated the need for social organizations to push for the inclusion of their members in the programme, and limited the scope for politicians to dole out transfer payments in return for votes. This has helped improve public opinion on transparency in programme management and the general absence of clientelist uses of programme funds IMPROVEMENTS IN DEVELOPMENT OUTCOMES Introducing democratic accountability mechanisms into social service delivery has helped improve both the take-up of public services and development outcomes. By reaching virtually all of Brazil s poorest families and helping to extend access to public services to them, the programme has produced tangible results in various areas of social policy, such as education, health and poverty reduction. According to Ferreira and Robalino (2010), CCTs in Latin America have helped increase school enrolment, bringing poor children to school, and increased the demand for health services among the poor. Targeting results also underscore the fact that the programme is reaching the poorest. Programme targeting reaching the poor Very good targeting results show that the programme is indeed reaching its intended targets the poorest and most vulnerable families (Cunha and Câmara 2008, Lindert et al. 2007). Household survey data from 2004 show that the BFP was already extremely well targeted, with the poorest population quartile receiving 80 per cent of programme transfers. Of the remaining 20 per cent, the data show that 85 per cent of these leakages went to families in the next poorest quartile, which could be considered near poor. The BFP shows far better targeting results than the cash transfer programmes it replaced, and also far out-performs public social insurance programmes such as social security and unemployment insurance which are highly regressive. The BFP performs very well in international comparisons. Data show 75 per cent of BFP benefits accruing to the poorest quintile, while the poorest 40 per cent of the population received 94 per cent of programme transfers. These results put the BFP at the top of the ranking in terms of the targeting accuracy of CCTs and social assistance programmes in Latin America, and among the top six of 85 transfer programmes in the developing world (Lindert et al. 2007). Moreover, the Cadastro Único has undergone significant reforms and improvements since the 2004 household surveys were carried out, which should have strengthened targeting outcomes even further. 17